DISTRIBUTION NETWORK RELIABILITY ENHANCEMENT THROUGH RELIABILITY BASED METHODOLOGY
Thando Roselette Khumalo (Montso)*; Jan Harm Christiaan Pretorius**
*Master of Engineering (Engineering Management) student, Faculty of Engineering and the Built Environment, University of Johannesburg, Johannesburg, South Africa
** Professor: Faculty of Engineering and the Built Environment University of Johannesburg, University of Johannesburg, Johannesburg, South Africa.
Abstract - Looking at the economic and technological growth of South Africa over the last 50 years up to 2016, it is evident that electricity is an essential factor. The need for a reliable power distribution system has grown tremendously over this period. The majority of households and businesses rely on the functions of electric appliances and equipment to assist with daily tasks such as cooking, bathing, etc.
Customer and energy based reliability indices are important to evaluate a predictive performance of distribution network systems. A case study of energy demands in Jabulani, Soweto, has been presented to evaluate the optimum failure rate and repair time for each section, so as to achieve the desired reliability goals in terms of the mentioned indices. Applying this reliability method helps in monitoring the distribution network feeders and, consequently, improving their reliability.
This paper describes a methodology for reliability enhancement of distribution network system by calculating the reliability indices such as (SAIDI-) System Average Interruption Duration index, (SAIFI-) System Average Interruption Frequency Index and (RSLI-) Reticulation System Loss Index, in order to determine the improvements that can be made to the distribution network using the reliability based methodology.
Index terms: MV – Medium Voltage, kV – kilovolt
COUE-Cost of unserved energy, CNC- Customer Network Centre, PRF-Project Review Forum
I. INTRODUCTION
Electricity is an essential product in South Africa. Considering the rate at which the country is growing, it has actually become more essential than ever as many households depend on it to carry out their day to day tasks. Reliability has always been a concern in the energy sector, but concerns are escalating as energy demand increases, and the political stability of many energy supply regions becomes more questionable (1). The provision of reliable, secure and affordable energy services is central to addressing many of today’s global development challenges. This includes poverty, inequality, climate change, food, security, health and education (2). This energy is also required for wealth creation and economic development. The South African bill of rights indicates that every citizen in the country has the right to electricity.
In the White Paper on the Energy Policy of the Republic of South Africa (December, 1998) from the Department of Minerals and Energy, the following definition is provided regarding Universal Access:
“Government recognises that household access to adequate energy services for cooking, heating, lighting and communication is a basic need. Whilst these needs can be met by various fuel-appliance combinations, Government recognises that without access to electricity, a clean, convenient and desirable fuel, human development potential is ultimately constrained. Government commits itself to implementing reasonable legislative and other measures, within its available resources, to progressively realise universal household access to electricity.
Access to electricity is taken to include grid supplies, Solar Home Systems, generators, hybrid systems, battery systems or any other supply solution which provides an appropriate and affordable electricity supply. The decision of which technology to utilise, will be based on life cycle cost analysis and the number of connections made in terms of the budget allocation.”
Currently Eskom has more than 8,000 individual Medium Voltage (MV) feeders (i.e. 1 kV – 33 kV) supplying more than 4.5 million customers, most of which were designed on a least-capital-cost basis with limited redundancy and back-feed capability. Focused initiatives are underway to improve the performance of the MV network in order to improve Eskom”s Distribution overall network reliability and technical performance (3).
II. PROBLEM STATEMENT
A. Problem Statement
The problem with Eskom’s Distribution is maintaining a reliable electricity delivery network system at a reasonable cost (4). Most of the functions at Eskom’s Distribution still
exclude cost effective, reliability management in their early planning stages (functions such as procurement), and this further limits strategic sourcing opportunities in leveraging the overall saving for Eskom.
The biggest challenge facing Eskom is the imbalance of supply and demand of electricity. As the current Government is trying to electrify many of the households previously excluded during the apartheid era, the country is at risk of electricity load shedding from time to time.
Throughout the lifecycle of the power distribution network, zero interruptions and faults are ideal, but different factors impact upon the performance of the network. The causes are mainly due to the design and topology of the network, planned operation and maintenance sessions, but these are not limited to physical infrastructure failures.
The focus is on the factors that contribute to power outages such as unplanned faults and failures that occur on the physical structure of the network, of which the typical causes are: •Incorrect design application
•Poor construction •Equipment overloading
•Poor conditions due to inadequate maintenance/ refurbishment
•Environmental factors (lightning, vegetation, and pollution). All of these have an impact on the undesireable load shedding situation.
This situation has a direct impact on the public protest that the residents of Soweto have embarked upon.
Many countries have faced challenges similar to those of the South African power utility (2). On the one hand, the recent multi-day blackouts across much of India made it painfully clear that the ability to correctly predict (and react
quickly to) power disaster scenarios is prized by risk, adverse shareholders, sympathy-infused publics, and politicians keenly attuned to the 24-h news cycle (5).
On the other hand, India’s blackouts also demonstrated that resources – financial, technical, and ecosystemic – are in perennially short supply, engendering a focus on the primacy of ephemeral successes at the expense of long-term sustainability, investment, and durability. The different best practices revived the distribution grid of the countries such as Brazil, Uganda, UK and India.
B. Objective
Reliability management is one dimension which can contribute to the bottom line at Eskom. The impact of reliability management has a positive outcome not only for Eskom’s profit, but also for the South African economy and the society at large. The research conducted now, hoped to find solutions to current power reliability problems at Eskom, and to provide Eskom’s management, (particularly at Eskom Distribution), solutions and develop preventative measures to the power delivery problems caused by an unreliable network in the country. This applies especially in political sensitive areas such as the biggest township in the country –SOWETO- as looked after by Eskom.
An efficient and reliable model at the network distribution planning and design stage of the network is necessary for monitoring this situation.
The case study in this paper will indicate the analyses and evaluation of the distribution network performance in Soweto, Jabulani, and to measure the effectiveness of implementing
Reliability Based Methodology in improving the performance of the future network.
III. METHODS
Publicly owned utilities typically attempt to design and operate electrical (generation, transmission and distribution) infrastructure in such a way so as to minimize failure in the grid, and cost to the public (6).Although load shedding is used to control the congestion in the network and to avoid severe power failures (7), the main objective is to have a continuous reliable power supply (without load shedding) to avoid negatively impacting the economy (8) .
The connection of customers to the South African electricity grid is governed by the National Energy Regulator of South Africa (NERSA) via the South African Grid Code developed from June, 2005, to August, 2007, (3). This code not only contains connection conditions but also investment criteria for such connections and is applicable not only to Eskom but also to other South African distributors, such as municipalities.
According to the grid code, a balance needs to be met between infrastructure cost options (both in terms of minimum cost of the energy supplied, as well as the customer interruption cost) and the network’s technical performance levels (RSA Grid Code, 2007, p9). The code highlights comprehensive energy systems planning, which aims at ensuring that energy-related policy and investment decisions consider all possible energy supply and demand side options, and are consistent with broader national goals (6). The methodology’s basic analytical approach used is by making use of modelled performance levels to inform decision making regarding network expected designed performance levels. These can provide a basis for better-informed
decision making (as opposed to, for example, historical or comparative international benchmarks) regarding the performance of the network (9). This is an approach that the Eskom Reliability Research Team is continuously researching in order to ensure that the network is performing at peak levels.
The main problem in the utility lies within the urban networks, where there is always growth in the notified demand. Thus, a qualitative research strategy was adopted in this research study since it involves the subject of an objective nature. Exploratory research had to be conducted as the subject is of a qualitative nature because of the limited volume of knowledge available on the subject considered (10).
The most obvious reason for performing a reliability prediction research methodology is to determine whether or not a proposed network design will meet a numerical, or other equivalent statement of reliability requirement that will improve the performance of the network (4).
Reliability network predictions were used to assess alternatives and to provide design direction in the early stages of network development and improvement (4). This could save time, money, other resources (such as personnel and transport), and it could improve the design .This method aids in the final network design stage where changes will not be very expensive.
In order to simulate and measure the network performance, the performance data was gathered using data loggers from the field, in order to analyse the electricity faults that were experienced in the sampled area. The data collected was analysed so that an informed decision could be made with regard to the enhancement of the concerned distribution
network. All faults, frequency of fault occurrences and times of the customer outages due to faults that occurred between January 2015 and January 2016 were recorded in the Network Enhancement Performance System (NEPS) which is used to indicate the performance of the feeders in the respective areas.
When the service of providing electricity comes to a halt, most processes stop immediately, or with a minimal delay, being contingent on other processes (9). This accumulation is evident in the valuation of uninterrupted electricity. The value of customer interruption costs was evaluated (9). Combining financial costs and potential Cost of Un-served Energy (CoUE) implications as well as broader economic and socio-economic impacts, and contrasting these costs with performance levels obtained, serves better to inform the reliability investment decision in a more holistic way as indicated Figure 1.
The workflow does consider the topology and design of the network as it has an inherent influence on the performance of the network which, in turn, impacts on the CoUE. For example: Feeder A - 11kVA feeder that is 10km long with 10 customers connected. Feeder B - 11kVA feeder that is 60km long with 10 customers connected. Assuming that the equipment, maintenance time, number of customers and interruption percentages are identical.
Sustained failure rate is calculated within the first stage of the network
⋋s = 0.19 per km per annum Total sustained failure rate is:
⋋s = 0.19 per km per annum * the total number of customer on the feeder.
Calculation for feeder A: Total ⋋s = 0.19 *10 = 1.9
Calculation for feeder B: Total ⋋s = 0.19 *60 = 11.4
It is clear from these calculations that the longer the overhead feeder, the higher the expected frequency of outages on Feeder B than in Feeder A. This failure rate will be simulated in software called DigSilent©.
The decision made from using the below work flow will determine the reliability measure that will need to be implemented in the distribution network. Thus, the performance indices that indicate the performance of the network are then analysed.
Fig 1: Reliability modelling decision work flow
C. Determination of performance levels for an electricity distribution network. Reliability indices are used to determine the level of performance of the distribution power network. These distribution reliability indices reflect the ability of sustaining service of power supply and the level of customer satisfaction (11). Traditional distribution reliability indices used in this study only represent the impact of interruptions on customers, especially those interruptions longer than 1minute (11).
In the context of this paper specific electrical network interruption performance indices applied are currently monitored by NERSA and incorporated in the performance compacting of utilities in South Africa. These specific indices are:
• SAIFI – Supply Average Interruption Frequency Index (12)
SAIFI is a measure of how often a customer would experience sustained interruptions on average for a measurement period, typically a supply period of a year. SAIFI can be calculated as:
∑
Served customers Connected Total expressed as interruptions per year.
• SAIDI – Supply Average Interruption Duration Index
SAIDI is a measure of how long a customer would experience sustained interruptions on average for a measurement period, typically a supply period of a year. SAIDI can be calculated as:
$ ∑
Served customers Connected Total expressed as hours per year.
• CAIDI – Average sustained interruption duration a customer would experience
Customer Average Interruption Duration Index is a measure of how long an average
interruption lasts for a measurement period, typically a supply period of a year. CAIDI can be calculated as:
∑
Total Number of Customer Interuption expressed as hours per interruption
CAIDI can also be expressed as a function of SAIDI and SAIFI:
$ $
SAIFI
• RSLI – Reticulation Supply Loss Index RSLI is a measure of how long the capacity of the system on average was interrupted for a measurement period, typically a supply period of a year.
RSLI ∑ kVA Hours Lost Inturruptions Total Connected kVA Served expressed as hours per year.
For illustrative purposes the analysis used focuses mainly on SAIDI (although our modelling also calculates SAIFI, CAIDI and RSLI). In order effectively to analyse and meaningfully understand reliability as a factor that influences design of the distribution network, the measurement and calculation of these indices is crucial (13) .
IV. RESULTS
The case study chosen for the purpose of this study is a project that is done in a very politically sensitive area. Soweto is the largest Township situated to the south of Johannesburg, Gauteng. Figure 1 indicates the Soweto HV/MV power network.
Fig 2: Future network indicating the proposed Jabulani 132/11kV substation
The community has been experiencing a number of power outages in the Jabulani area. .The power outages experienced are a result of many factors such as: illegal connections, non-payment of electricity, constrained network and theft of infrastructure, to mention but a few. Reliability based Methodology was implemented in this project to measure and analyse the performance of the distribution Network.
The research results from the data collected from the field were calculated and presented below through the NEPS* software:
The Performance indices results were as follows for the period starting in 01 January, 2015, until 31 January, 2016.
SAIDI: The SAIDI target for period January, 2015, to January, 2016, was marked or capped at 14.61 hours. This represents the acceptable number of hours in a year that the customer can be without power .The Jabulani substation is not in operation yet, thus results indicate that the Zola CNC failed to meet the target, since the highest number of hours was 30.36.
Fig 3: SAIDI for the Zola CNC
SAIFI: The SAIFI target for the period January, 2015, to January, 2016, was marked or capped at 6 power interruptions/customer. The results indicate that the average frequency of 6 interruptions/customer was not realised, since the highest average frequency was 7, 07.
Fig 4: SAIFI for the ZOLA CNC
RSLI: The SAIFI target for the period January, 2015, to January, 2016, was marked or capped at 2.68 per incident of supply loss. Results show that the target was not reached since SLI was at 2.74. Thus the performance of the distribution network in Jabulani Soweto is almost within the stipulated limits.
Figure 5: RSLI for Zola CNC
The were three root causes for the worst events that impacted on the SAIDI for the financial year 2015/16 in the area were as follows: 57.24% was due to a cable fault, as the cable was overloaded because of illegal connections, 38.27% was due to a storm that occurred during November, 2015, and
4.49% was due to a fuse failure in one of the transformers.
Figure 6: Power interruptions root cause for the Zola CNC
An informed decision made was to plan and design a new substation named Jabulani 132/11kV substation. This new substation proposed to alleviate the power constraints in the area and to supply the new Jabulani precinct. The plan to build it by 2018 was approved by the Eskom Gauteng (PRF) committee.
The aim of the study was to research the effectiveness of implementing reliability based methodology to enhance the distribution network system by calculating the reliability indices such as SIADI, SAIFI and RSLI in order to determine the improvements that can be made to the distribution network in Soweto Jabulani.
This was achieved by assessing the worst contributing events to a distribution network by calculating the identified indices, reviews on existing literature on distribution reliability and assessing the existing infrastructure. The effectiveness was proven through a load forecast that indicated that 80% of the feeders and substation transformers would not be overloaded, thus performance would be significantly improved once the new station is in operation.
Customer and energy based reliability indices are important in evaluating predictive performance of distribution systems. An application of the reliability model has been presented to evaluate the optimum failure rate and repair time for each section so as to achieve the desired reliability goals, in terms of the mentioned indices. In the distribution network the risk of having supply interruption goes further on impacting the economy at large, thus reliability engineering, when networks are planned and designed, has to be implemented.
The conclusion is that the topology might need to be adjusted to improve reliability and enhance its performance.
Fig 7: Current Taunus network topology
Fig 8: Future Taunus network topology
The research indicated that adverse weather conditions are also contributing factors in the bad performance of the distribution network. Thus, faults can never be 100% eliminated since adverse weather conditions are beyond human control. The visibility of the cost of unserved energy to the customer will indicate
largely the impact on the economy if power is not delivered at the right time of need to the customer. The reliability based model will help solve one of the most critical performance criteria for the utility.
This study has also illustrated that the utilization and implementation of reliability based methodology is very effective when developing the network, since it improves the performance of the distribution network and results in a continuous supply of power. Consequently, customers will be happy and the cost of unserved energy will be significantly reduced.
Further research can be done on the Network distribution enhancement in areas such as Diepsloot, North of Johannesburg, to alleviate power constraints that exist and enhance the distribution network performance.
IV. REFERENCES
[1] Ryan W. McCarthy, Joan M. Ogden and Daniel L. Speling: “Assessing reliability in energy supply systems”. Energy Policy, Vol. 35, Carlifonia , 2007.
[2] Brian Ó Gallachóir and Morgan Bazilian:“Energy access scenarios to 2030 for the power sector in sub-Saharan Africa”. Utilities Policy, pp. 1-16, 2012.
[3] D Gütschow and Dr C Carter-Brown: “Quantifying the Reliability Impact of Smart Grid Technologies on Medium Voltage Overhead Networks”. CIGRE , 7th
SOUTHERN AFRICA REGIONAL
CONFERENCE, 2013.
[4] Martin Cameron and Dr Clinton Carter-Brown: “Electrical Utility distribution network capital planning - a network reliability informed approach to prioritising investment for economic sustainability”. Vol 1, AMEU, 2012.
[5] R.E.Brown: “Electric Power Distribution Reliability, Second edition”. ISBN-13:9780849375675, CRC Press , USA, p. 504 , 2008.
[6] Yi Ding , Peng Wang , Lalit Goel , Roy Billinton and Rajesh Karki: “Techniques, Reliability assessment of restructured power systems using reliability network equivalent and pseudo-sequential simulation”. 1665–1671, Singapore : Science Drirect, Vol. 77, 2007.
[7] B. Yssaad, M. Khiat, and A Chaker: “Reliability centered maintenance optimization for power distribution systems”. Algeria : Electric Power and Energy Systems, Vol. 55 , 2014.
[8] Osmo Siirtoa and Markku Hyvärinena: “Improving reliability in an urban network”. issue120, Helen Electricity Network Ltd , Finland 2015.
[9] Partington, David: “Essential Skills for Management Research”. SAGE Publications Ltd, London , 2002.
[10] Xiao Xiangning, Tao Shun, Bi Tianshuand Xu Yonghai: “Study on Distribution Reliability Considering Voltage Sags and Acceptable
Indices”. Beijing : IEEE TRANSACTIONS ON POWER DELIVERY, Vol. 22 2007.
[11] Richard E. Brown: “Electric Power Distribution reliability”. Marcel Dekker Inc, New York . ISBN: 0-8247-0798-2, 2002 .
[12] T.B. Mavuso, Prof. J.H.C. Pretorius and Dr. A. Wessels: “ Reliability Based Planning Methodology for Distribution Feeder Automation”. Johannesburg : University of Johannesburg, 2015.
